hello we want to ask for advice because there are problems related to the use of the signoz application in production. When opentelemetry on the application server enabled sending trace data, memory and cpu utilities on the signoz server rise significantly and cause the signoz server to hang. this utilities spike occurs when the application operating hours start running 6AM - 7 PM Weekday. Previously did not happen in the development environment. what should we do?
cc: @Mohamad Akmal Faudzan@ksltch
s
Srikanth Chekuri
12/10/2024, 3:13 PM
Resource usage will be directly proportional to the amount of data you are sending; in development, it might be very small data you tested with. What are the resources available and how many traces/logs/metrics are you ingesting?
m
Muhammad Fajar Pratama
12/12/2024, 2:09 AM
The data received by Signoz is the Odoo Apps log file, Odoo server metrics and queries to PostgresQL along with queries per second. we are trying a batching system for sending data from opentelemetry on odoo to signoz. Do you recommend this method? and one more thing, is upgrading the CPU and memory a permanent solution to solve this problem? we are currently using minimum specifications (cpu 4 core ram 8gb)
s
Srikanth Chekuri
12/12/2024, 6:14 AM
You haven't shared the actual volume of data
m
Muhammad Fajar Pratama
12/13/2024, 8:25 AM
in peak hour we have got 1.4k Ops along with the increase in CPU and RAM and finally server down. I also added the signoz data that was already saved.
notes: the trace is not currently running. and it only ran for less than 5 days because when the trace was run, a few hours later the server died
cc: @ksltch@Mohamad Akmal Faudzan
s
Srikanth Chekuri
12/13/2024, 2:39 PM
You don't have a lot of data, but the sudden peaks are problematic.